Worsening tricuspid regurgitation after ICD implantation is rather due to transvenous lead than natural progression

BACKGROUND

Transvenous implantable cardioverter-defibrillators (TV-ICDs) are associated with greater tricuspid regurgitation (TR) severity, which leads to increased mortality. The pathophysiology is assumed to be lead-related, hence, treatment includes lead extraction. However, TR may also naturally occur in the high-risk ICD population, or may be caused by right ventricular pacing. We sought to evaluate the effect of ICD type (with or without lead) and pacing percentage on post-implantation TR severity.

METHODS

In this retrospective cohort study, consecutive patients were included with a primary S-ICD or TV-ICD implantation between 2009 and 2019 and echocardiography studies ≤3 months before and ≤ 3 years post-implantation. The effect of ICD type on TR severity at follow-up was estimated adjusting for ventricular pacing percentage and potential confounders. The effect of ventricular pacing percentage on TR severity at follow-up was adjusted for potential confounders.

RESULTS

118 patients were included (mean age 52 ± 21): 31 (26%) with an S-ICD and 87 (74%) with a TV-ICD. Median 20 months post-implantation, worsening TR was found in 11/31 (34%) S-ICD patients and 45/87 (52%) TV-ICD patients (p = 0.15). Adjusted for age, atrial fibrillation, baseline TR and mitral regurgitation, ventricular pacing percentage, ICD dwelling time, BMI, hypertension and left ventricular ejection fraction, TV-ICDs were significantly associated with greater TR severity (OR 9.90, p = 0.002). Ventricular pacing percentage was very low, and not significantly associated with greater TR severity (OR 0.95, p = 0.066).

CONCLUSIONS

Our results suggest that greater TR severity in ICD patients is mainly caused by the transvenous lead, rather than natural progression in the ICD population.

Real-world long-term battery longevity of leadless pacemakers

Background

Leadless pacemakers (LPs) have proven safe on the mid-term. Long-term safety will largely depend on the need for replacements with a concomitant risk of complications. As replacements are mainly due to battery depletion, battery longevity is, in essence, an important determinant of long-term safety. Mean battery longevity of the most often implanted LP (Micra VR) predicted by the manufacturer is 12 years, but mid- and long-term data is lacking.

Purpose

We sought to determine the long-term battery longevity of this LP in a real-world cohort.

Methods

Consecutive patients with an LP implantation at our tertiary hospital from January 2014 to September 2021 were included in a prospective cohort. Baseline characteristics and electrical parameters during all yearly follow-up visits were assessed.

Results

105 patients (73 [70%] male, age at implantation 80±9 years) were included. Pacing indications were atrial fibrillation with slow conduction (34 [32%]), third degree AV block (35 [33%]), incomplete AV block (12 [11%]), sinus node disease (23 [22%]) and unexplained syncope (1 [1%]). At implantation, pacing threshold, impedance and R-wave amplitude were 0.76±0.64V at 0.24ms, 790±220Ω and 11.1±4.9mV, respectively, lower rate was programmed at 50–60/min in 88 (84%), <50/min in 8 (7.6%) and >60/min in 9 (8.6%) and rate-responsive pacing was programmed on in 51 (49%). Median follow-up was 25 months (range 2–88 months). Pacing percentage was 46% (IQR 12–98%) at year 0; 64% (10–99%) at year 5 and 91% (56–100%) at year 7.

No battery failures were seen during follow-up. At 3 years, expected battery longevity was >8 years (maximum value) in nearly all patients (49/50 [98%]) and at 5 years in the majority (16/18 [89%]) of patients; at 7 years, expected battery longevity ranged from 4.1 to >8 years (Figure). This implies an expected battery longevity of >13 years in the majority of patients. In two subjects, expected battery longevity was below the maximum value within 5 years after implantation. In the first, latest follow-up visit was after 3 years and expected battery longevity was 3.0 years. An increase in pacing percentage of +70 percent points was seen during follow-up and there was a relatively high pacing capture threshold (1.5V@0.4ms). In the second, the latest follow-up visit was at 6 years, and expected battery longevity was 1.9 years. This was caused by a large increase in pacing percentage (+80 percentage points) and lower rate of 70/min; pacing threshold was 0.5V@0.24ms. We did not see clear overall differences in expected battery longevity at 5 years, when the patients were stratified per quartile pacing percentage. Of 34 deaths during follow-up, 31 (91%) had expected battery longevity >8 years at the last follow-up visit.

Conclusions

These long-term real-world data suggest that the battery longevity of the Micra LP will exceed the expected battery longevity. Data from larger registries is necessary to confirm these results.

Funding Acknowledgement

Type of funding sources: None.

Reduction of artifacts and noise in small electrogram datasets without manual annotation using transfer machine learning

Background/Introduction

Mapping AF is challenging. Monophasic action potentials (MAPs) show that most of the recorded signals are not representing the atrial activity, and arise from far-field or other artifacts. Thus, reducing noise in electrophysiological signals is essential, yet it can be difficult for cross-talk from other chambers and pacing. Strategies to reduce noise include template matching, averaging, and smoothing, but all of them have major limitations. Furthermore, expert interpretation requires knowledge to discriminate signals from noise, but is subjective.

Purpose

We hypothesised a) that atrial and ventricular electrograms with varying artifact and noise can be denoised using autoencoder neural networks (NNs) without requiring manual annotation and in a reproducible manner, and b) we could train these NNs on a large available dataset ventricular signals, then apply transfer learning to the original smaller atrial dataset. We applied this approach to MAPs, which have interpretable shapes and would help identifying local from far-field signals helping in diagnosis, mapping and ablation.

Methods

We first trained with 5706 left and right ventricular MAPs from 42 patients with ischemic cardiomyopathy (age 65±13y; Fig. 1A), with 60% for training, 20% (validation) and 20% (testing). Transfer learning and parameter-tuning were then used to apply this NN to a smaller sample of atrial MAPs (N=641 from 21 patients, 67±5y, 13 women; Fig. 2B, D, F). The autoencoder was used to eliminate pacing artifacts in ventricular MAPs (Fig. 1B, C) and denoise atrial MAPs (Fig. 2C, E, G) by reconstructing key learned features. The accuracy of the reconstruction was evaluated using Pearson Correlation Coefficient (PCC) and a novel similarity coefficient (SC). No manual annotation was required to identify noisy signals.

Results

The trained NN encoder learned key features of ventricular MAPs and reconstructed these clean signals with a SC=0.91±0.16 and PCC=0.99±0.01 (Fig. 1A). With this training, the NN was able to denoise ventricular MAPs with pacing artifact (Fig. 1B, C). After fine-tuning, the NN learned key signal features (upstroke, triangular descent, terminus) and thus reduced diverse noise without specific training or manual annotation. Namely, it was able to reconstruct atrial MAPs eliminating ventricular noise, high frequency noise and truncated signals (Fig. 2).

Conclusions

Machine learned encoder-decoders are powerful tools that can learn essential features of atrial and ventricular signals and hence isolate noise. Transfer learning is effective when large datasets are unavailable for training. This approach can separate atrial beats from far-field ventricular beats and other sources of noise. The ability to eliminate a diverse range of noise improves this approach over existing techniques and may have far-reaching applications in electrophysiology, such as mapping and ablation.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NIH

Clinical phenotyping of implantable cardioverter defibrillator patients to identify the association of remote device monitoring on survival benefit: a cluster analysis

Background/Introduction

Over the past decade, various studies have demonstrated an association between remote device monitoring (RM) in implantable cardioverter-defibrillator (ICD) patients and all-cause mortality. However, it is unclear which clinical phenotypes account for this survival benefit.

Purpose

We aimed to identify clinical phenotypes within an ICD population on RM and conventional follow-up (NRM), and evaluate differences in long-term survival per clinical phenotype.

Methods

This is a single-center, retrospective, observational study of de novo ICD implantations (single- and dual chamber, biventricular and subcutaneous) between 2010 and 2021 in patients ≥18 years old. Data was extracted from electronic health records. Unsupervised two-step cluster analysis (TSC) was performed to identify distinct clinical phenotypes in the RM and NRM cohorts. Multinomial logistic regression analysis was performed to identify cluster characteristics, differences among the different clusters were analysed by Chi-squared test. Variable importance was evaluated for each of the 18 included clinical variables. Survival analysis was performed using the Kaplan-Meier method, log-rank was used to compare survival between groups.

Results

A total of 1872 ICD patients were analysed using TCS, comprising 1265 RM and 607 NRM patients, respectively. TCS for a mean follow-up duration of 5.4±3.1 years partitioned the RM and NRM groups into six clusters (Table 1). Clusters RM1 (n=444) and NRM3 (n=220) represented a predominantly primary prevention and DCM phenotype. Clusters RM2 (n=300) and NRM2 (n=173) indicated young and secondary prevention phenotypes. Clusters RM3 (n=220) and NRM1 (n=214) comprised an older, male and ischaemic cardiomyopathy (ICM) phenotype. In survival analysis 5-year cumulative incidence of mortality for the subsequent clusters was 7.4%, 4.5%, 11.6%, 41.4%, 29.8%, and 25.1%, respectively (log rank p-value <0.001, Figure 1). Thus, younger and DCM phenotypes (often with genetic aetiologies) experienced superior survival and high rates on RM, whereas older, male and ICM phenotypes with low rates on RM were associated with worse survival. In addition, between-group cluster-comparison of patients on RM only demonstrated important differences in long-term survival, where RM3 (comprising older, male with predominantly ICM) had poorer survival compared to young patients with a high prevalence of DCM, primary arrhythmia syndrome and low proportions of comorbidities (i.e. atrial arrhythmias, cerebral vascular accidents, diabetes mellitus).

Conclusion

We identify novel clinical phenotypes (clusters) of ICD patients on RM and having conventional follow-up with large differences in long-term survival. Future studies should further explore if these differences reflect benefit from RM, or the underlying natural history of each phenotype. These results may accordingly guide monitoring strategies tailored to specific ICD patient profiles.

Funding Acknowledgement

Type of funding sources: Public hospital(s). Main funding source(s): Public funding

Artificial intelligence to reduce artifact in cardiac electrophysiological signals

Background/Introduction

Signals in Electrophysiology cases are often noisy despite laboratory shielding and filtering, and current noise-reduction methods are suboptimal. Template matching can identify a “nearest type” of electrogram, but libraries of signal shapes may be unavailable. Beat averaging can reduce noise but obscures beat-to-beat variations and is not optimal to analyze dynamically changing signals, such as when moving a catheter in the heart. Smoothing reduces noise yet blurs high frequency components.

Purpose

We set out to test if machine learned autoencoders could reduce noise in single beats without requiring massive training data or beat libraries. Specifically, we hypothesised that noisy electrograms in small datasets of atrial signals could be de-noised using an encoder-decoder neural network (NN) using transfer learning of machines trained to recognize key features in larger datasets of related signals.

Methods

We applied NN to monophasic action potentials (MAPs), because they have visually verifiable shapes. The NN was first trained to reconstruct 5706 left and right ventricular MAPs in 42 patients (67±13y; Fig. 1A). Transfer learning was then used to apply the NN to a much smaller dataset of 641 atrial MAPs in 21 patients (67±5y, 13 women; Fig. 1B, D, F).

Results

NN reconstructed atrial MAPs with a Pearson correlation of 0.87±0.11. After fine-tuning, NN reconstruction accuracy improved dramatically (Pearson 0.99±0.01; p<0.001). In Fig. 1B–G the NN learned key MAP features (upstroke, triangular descent, terminus) and thus could eliminate ventricular artifact and electrical circuit noise without specific training or manual annotation.

Conclusion

Machine learned autoencoders are a novel and powerful approach to de-noise electrophysiological signals in a dynamic, beat-to-beat fashion. The ability to learn fundamental signal features from models trained in large datasets, and apply them via transfer learning to small datasets in different heart chambers may have wide ranging applications for automated signal annotation, mapping and ablation.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NIH

Remote device monitoring shows long-term survival benefit in a propensity matched study of ICD patients

Funding Acknowledgements

Type of funding sources: None.

Background

Remote monitoring (RM) for implantable cardioverter defibrillators (ICDs) enables early detection of clinical events and reduces inappropriate shocks. However, although RM has been associated with improved clinical outcomes at short to medium-term follow-up, data on long-term survival benefit of RM is currently lacking.

Purpose

To investigate long-term survival of patients using RM and compare this to a propensity score matched non-RM group.

Methods

This is a retrospective, observational single-center analysis of patients ≥18 years old implanted with an ICD (single- and dual chamber, biventricular or subcutaneous ICDs) between 1995 and 2021. Data was extracted from the electronic health records. Patients were included in the RM group only if they were started on RM at some point during follow-up. To adjust for differences in baseline characteristics between RM and non-RM patients, propensity score matching was performed in a 1:1 fashion (caliper 0.20) using greedy matching. Estimation of the propensity score was done using logistic regression with RM as the outcome variable, adjusting for 17 covariates (age, gender, year of device implant, type of device, clinical variables and medication at baseline). Time-to-event analysis was performed using Kaplan Meier survival analysis, with significance indicated using a log rank P value. Hazard ratios and 95% confidence intervals (CI) were calculated using a Cox proportional-hazards model.

Results

A total of 3199 ICD patients were included in this analysis. After propensity score matching for the probability of RM use, 1160 RM patients and 1160 non-RM patients with similar baseline characteristics were selected (mean age 59.1 ±14.8 years, 72.8% male). During a mean follow-up duration of 8.1 ±5.2 years, 363 (31.3%) patients died in the non-RM group and 111 (9.6%) patients died in the RM group (log rank p<0.001, Figure 1). The hazard ratio of RM for mortality was 0.261 (95 CI% 0.211 – 0.323, p=0.001).

Conclusion

Long-term retrospective analysis indicates a significant survival benefit in ICD patients using RM.

Remote device monitoring for implantable cardioverter defibrillators during the COVID-19 pandemic

Funding Acknowledgements

Type of funding sources: None.

Background

Remote monitoring (RM) for implantable cardioverter defibrillators (ICDs) is recommended as the standard of care in clinical guidelines. Presumably, the restrictions on face-to-face visits that were imposed during the coronavirus (COVID-19) pandemic have further accelerated the adoption of RM. However, quantitative real-world data on the uptake of RM during the COVID-19 pandemic is lacking.

Purpose

To assess the uptake of RM during the COVID-19 pandemic to a pre-COVID-19 period, and compare the arrhythmic burden between the two groups.

Methods

This is a substudy of the retrospective, observational single-center DISTANT-study. For this substudy, data from patients who were enrolled in the RM program after de novo ICD implantation (single- and dual chamber, biventricular or subcutaneous ICDs) were analysed. The time until RM was initiated per patient was calculated for patient implanted during the COVID-19 pandemic (March 2020-January 2021) and compared to a similar 10-month period pre-COVID-19 (May 2019-March 2020). ICD therapy (shock and/or anti-tachycardia pacing), non-sustained ventricular tachycardia (NSVT), supraventricular tachycardia (SVT) and mortality were registered for each patient. Patients <18 years old at implantation and patients with a follow-up of <6 months were excluded from this analysis.

Results

A total of 134 patients (72.4 % male, mean age 57.3 ±14.9 years) were eligible for this substudy, of which 61 patients in the COVID-19 group and 73 patients in the pre-COVID-19 group. In both groups there was a similar percentage of primary prevention ICD implantations (COVID-19: 43%, pre-COVID: 44%; p=0.888). During COVID-19, RM was initiated more promptly following ICD implantation compared to pre-COVID-19 (respectively 63 days vs. 131 days; p=0.007). Second, in the COVID-19 group 60.7% patients were enrolled in RM within 30 days following implantation compared to 39.7% in the pre-COVID-19 group (p=0.016). In terms of arrhythmic burden, no differences in the occurrence of ICD therapy (p=0.759), NSVT (p=0.267) and SVT (p=0.454) were observed.

Conclusion

During the COVID-19 pandemic RM was initiated more promptly following ICD implantation compared to before the pandemic, however, no differences in arrhythmic burden between groups were observed.

Pacemaker choice optimization based on expected pacing requirement

Funding Acknowledgements

Type of funding sources: None.

Background

Bradyarrhythmias are adequately treated with pacemakers. Currently, different pacing modes (single- and dual-chamber, cardiac resynchronization therapy and physiologic pacing) and types (leadless, transvenous) are available. Expected pacing requirement is an important factor in determining optimal pacing mode and type.

Purpose

To evaluate atrial (AP) and ventricular pacing percentages (VP) over time, stratified by pacemaker indication.

Methods

All consecutive patients ≥18 years with a DDD(R) transvenous pacemaker and ≥1 year follow-up at a tertiary center between January 2008 and January 2020 were retrospectively included. Baseline characteristics and AP and VP at yearly follow-up visits were assessed.

Results

714 patients were included, of which 210 (29%) with incomplete AV block (AVB), 261 (37%) complete AVB, 243 (34%) sinus node disease (SND). Mean age at implant was 72±13, 68±16, 67±14 years (p=0.005). AP and VP over time is shown in Figure. For incomplete AVB, median AP increased from 3% (1-24%) to 21% (1-49%) and VP from 37% (7-93%) to 91% (16-100%); for complete AVB, median AP increased from 3% (1-17%) to 10% (2-30%) and VP remained stable (98% [13-100%] to 99% [62-100%]); for SND, median AP increased from 26% (5-63%) to 55% (8-90%) and VP remained stable (3% [1-13%] to 5% [1-31%]).

Conclusion

These results confirm the pathophysiology of different pacemaker indications, resulting in clear differences in pacing requirement over time and expected battery longevity. These results can be used to help guide optimal pacing mode and suitability for leadless or physiologic pacing.

Leadless pacemakers as replacement for infected transvenous pacemakers: different strategies are feasible

Funding Acknowledgements

Type of funding sources: None.

Introduction

Pacemaker infections have a high morbidity and mortality and are an indication for extraction. For reimplantation, leadless pacemakers (LPs) may be preferable due to a low chance of infection. Even more, early LP reimplantation in pacemaker-dependent patients would circumvent the need for temporary pacemakers.

Methods

We included all patients with LP implantation before, simultaneously with, or after transvenous pacemaker extraction due to infection, between January 2013 and December 2021. Outcomes were assessed during standard follow-up visits.

Results

30 patients (mean age 81±8 years) were included, of which 19 (63%) had a pocket infection, 10 (33%) endocarditis and 1 (3%) a systemic infection without endocarditis (Table). LP implantation was successful in all and was performed before extraction in 2 patients (7%; 3 and 5 days before), simultaneously in 6 (20%) and after extraction in 22 (73%). There were 3 procedural complications: 2 femoral artery bleedings and one LP dislocation. Also, one patient with complete AV block and an initially stable escape rhythm had an in-hospital cardiac arrest due to asystole after transvenous pacemaker extraction, but before LP implantation. During follow-up of median 21 months (IQR 7-53 months), no reinfection occurred. Six Nanostims were extracted due to early battery depletion, prophylactically after the battery advisory, or due to non-capture (median 36 months [range 0-67 months] after implantation); histopathologic examination of tissues around the devices showed no signs of infection. Two Nanostims were abandoned after which another device was implanted. No further device revisions were necessary during follow-up.

Conclusions

In case of transvenous pacemaker infection, LP implantation before, simultaneously with or after extraction is safe and effective.

Adherence to wearables in implantable cardioverter-defibrillator patients: Preliminary results from the prospective, multicenter SafeHeart-study

Funding Acknowledgements

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): Horizon2020

Introduction

Wearable devices are gaining interest in the clinical assessment of physical behavior as a marker of disease severity. With the increased use, patient willingness and adherence will be increasingly important. As part of the SafeHeart study, examining the potential of physical behavior as an identifier of clinical deterioration in patients with an implantable cardioverter defibrillator (ICD), we present preliminary results on adherence to a wrist-worn wearable used for physical behavior assessment.

Purpose

Define the willingness to participate and long-term adherence to wearables in an ICD population.

Methods

This is a preliminary analysis of the ongoing multicenter, prospective, observational SafeHeart study. SafeHeart is aimed to construct a personalized prediction engine for ICD therapy using wearable-assessed physical behavior, remote ICD monitoring, electronic health records, and patient-reported data. The study will enroll 400 participants with an ICD with or without cardiac resynchronization therapy (CRT-D). In this preliminary analysis, wearable data was analyzed for the first 50 participants, where inclusion required a minimum of 1 month of follow up data. No data from the wearables were provided to the participants. The wrist-worn wearables were used continuously (day and night) for up to 12 months of follow-up. Adherence to the wearable was measured through patient-reported (subjective) adherence and wearable-measured (objective) adherence. Data were extracted from the wearables and non-wear time was detected via open source algorithms. A valid day was set to 22 hours of available wear time with 24-hour periods assessed from 3pm to 3pm for sleep metric capture. The willingness to participate and dropout rates were calculated for the same first 50 patients of the study.

Results

A total of 50 ICD participants were included in this study. The mean age was 65.1 years, 82 % male, with a mean follow up of 7 weeks, generating 326 patient weeks of data. Regarding patient-reported adherence, participants reported 81.4% full adherence and 18.6 % of participants reported very brief non-wear due to e.g. sauna or surgery. Of those reporting non-wear, 62.5% described one episode only of non-wear lasting 15-75 minutes. Regarding objectively measured adherence from wearable data, full adherence was shown in 91.7% of days. The mean number of valid days per participant was 41.3. Recruitment rates showed a willingness to participate of 50% (50/100) out of eligible subjects invited. No participants were lost to follow

Conclusion

Results show high adherence and reasonable willingness to participate without wearable adherence dropping over time. Comparison of objectively measured and patient-reported adherence showed similar values.

Electrocardiogram-based mortality prediction in patients with COVID-19 using machine learning

Background and purpose

The electrocardiogram (ECG) is frequently obtained in the work-up of COVID-19 patients. So far, no study has evaluated whether ECG-based machine learning models have added value to predict in-hospital mortality specifically in COVID-19 patients.

Methods

Using data from the CAPACITY-COVID registry, we studied 882 patients admitted with COVID-19 across seven hospitals in the Netherlands. Raw format 12-lead ECGs recorded within 72 h of admission were studied. With data from five hospitals (n = 634), three models were developed: (a) a logistic regression baseline model using age and sex, (b) a least absolute shrinkage and selection operator (LASSO) model using age, sex and human annotated ECG features, and (c) a pre-trained deep neural network (DNN) using age, sex and the raw ECG waveforms. Data from two hospitals (n = 248) was used for external validation.

Results

Performances for models a, b and c were comparable with an area under the receiver operating curve of 0.73 (95% confidence interval [CI] 0.65–0.79), 0.76 (95% CI 0.68–0.82) and 0.77 (95% CI 0.70–0.83) respectively. Predictors of mortality in the LASSO model were age, low QRS voltage, ST depression, premature atrial complexes, sex, increased ventricular rate, and right bundle branch block.

Conclusion

This study shows that the ECG-based prediction models could be helpful for the initial risk stratification of patients diagnosed with COVID-19, and that several ECG abnormalities are associated with in-hospital all-cause mortality of COVID-19 patients. Moreover, this proof-of-principle study shows that the use of pre-trained DNNs for ECG analysis does not underperform compared with time-consuming manual annotation of ECG features.

Supplementary Information

The online version of this article (10.1007/s12471-022-01670-2) contains supplementary material, which is available to authorized users.

Clinical presentation, disease course, and outcome of COVID-19 in hospitalized patients with and without pre-existing cardiac disease: a cohort study across 18 countries

AIMS

Patients with cardiac disease are considered high risk for poor outcomes following hospitalization with COVID-19. The primary aim of this study was to evaluate heterogeneity in associations between various heart disease subtypes and in-hospital mortality.

METHODS AND RESULTS

We used data from the CAPACITY-COVID registry and LEOSS study. Multivariable Poisson regression models were fitted to assess the association between different types of pre-existing heart disease and in-hospital mortality. A total of 16 511 patients with COVID-19 were included (21.1% aged 66-75 years; 40.2% female) and 31.5% had a history of heart disease. Patients with heart disease were older, predominantly male, and often had other comorbid conditions when compared with those without. Mortality was higher in patients with cardiac disease (29.7%; n = 1545 vs. 15.9%; n = 1797). However, following multivariable adjustment, this difference was not significant [adjusted risk ratio (aRR) 1.08, 95% confidence interval (CI) 1.02-1.15; P = 0.12 (corrected for multiple testing)]. Associations with in-hospital mortality by heart disease subtypes differed considerably, with the strongest association for heart failure (aRR 1.19, 95% CI 1.10-1.30; P < 0.018) particularly for severe (New York Heart Association class III/IV) heart failure (aRR 1.41, 95% CI 1.20-1.64; P < 0.018). None of the other heart disease subtypes, including ischaemic heart disease, remained significant after multivariable adjustment. Serious cardiac complications were diagnosed in <1% of patients.

CONCLUSION

Considerable heterogeneity exists in the strength of association between heart disease subtypes and in-hospital mortality. Of all patients with heart disease, those with heart failure are at greatest risk of death when hospitalized with COVID-19. Serious cardiac complications are rare during hospitalization.

Accelerometry-assessed physical behaviour and the association with clinical outcomes in implantable cardioverter-defibrillator recipients: a systematic review

Funding Acknowledgements

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): Eurostars

Introduction

Patients at a high risk of sudden cardiac death (SCD) benefit from an implantable cardioverter defibrillator (ICD). However, they remain at a high risk of (inappropriate) shocks, heart failure, mortality and psychological distress. Consumer-level wearable accelerometry as method for recording physical behaviour (PB) has gained popularity over the past years, but so far the clinical potential is largely underinvestigated. The identification of patterns in PB and the association with clinical outcomes may provide a means to improve ICD therapy.

Purpose

This review addresses the evidence concerning PB in ICD patients and aims to characterise PB patterns associated with clinical outcomes.

Methods

A systematic review of studies focussing on accelerometer-assessed PB in patients older than 18 years equipped with an ICD, or patients at a high risk of SCD (e.g. advanced heart failure) was performed. PB could be assessed using a wearable accelerometer or an embedded accelerometer in the ICD (i.e. device-measured physical activity (D-PA)). Papers presenting quantitative data in English language peer reviewed journals published between January 2000 and September 2020 were identified via the OVID MEDLINE and OVID EMBASE databases. A study protocol describing study selection, data charting and summarisation of results was developed apriori. Study selection was conducted by two independent reviewers and a third reviewer in case of disagreement.

Results

A total of 4219 studies were identified, of which 51 were deemed appropriate for this review. Of these studies, 29 examined D-PA (n = 169.742 patients), 19 examined wearable accelerometery (n = 1.601) and 3 validated wearable accelerometry against D-PA (n = 106). The main findings were that (i) a low level of physical activity (PA) after implantation of the ICD and (ii) a decline in physical activity were both associated with an increased risk of ICD shocks, hospitalization and mortality. Second, PB was affected by cardiac factors (e.g. onset of atrial arrhythmias, ICD shocks) and non-cardiac factors (e.g. seasonal differences, pandemic lockdown). Third, PB was related to left ventricular ejection fraction, physical and cognitive function and quality of life. The evidence regarding wearable accelerometry compared to D-PA was scarce and heterogeneous.

Conclusion

This review demonstrated the potential of PB as an identifier of clinical deterioration in an ICD population. Accelerometer-assessed PB data could improve early warning systems and facilitate preventive and pro-active strategies, especially considering the nature of PB as modifiable risk factor. We suggest two directions for future research: (i) prospective collection of wearable accelerometry data in an ICD population to identify the most clinically relevant behavioural metrics (ii) investigation of preventive measures that can be undertaken once changes in PB are observed. Abstract Figure. Accelerometry-derived physical behaviour

Electrocardiogram-based mortality prediction in patients with COVID-19 using machine learning

Funding Acknowledgements

Type of funding sources: Public hospital(s). Main funding source(s): The Netherlands Organisation for Health Research and Development (ZonMw) University of Amsterdam Research Priority Area Medical Integromics

OnBehalf

CAPACITY-COVID19 Registry

Background

The electrocardiogram (ECG) is an easy to assess, widely available and inexpensive tool that is frequently used during the work-up of hospitalized COVID-19 patients. So far, no study has been conducted to evaluate if ECG-based machine learning models are able to predict all-cause in-hospital mortality in COVID-19 patients.

Purpose

With this study, we aim to evaluate the value of using the ECG to predict in-hospital all-cause mortality of COVID-19 patients by analyzing the ECG at hospital admission, comparing a logistic regression based approach and a DNN based approach. Secondly, we aim to identify specific ECG features associated with mortality in patients diagnosed with COVID-19. 

Methods and results

We studied 882 patients admitted with COVID-19 across seven hospitals in the Netherlands. Raw-format 12-lead ECGs recorded after admission (<72 hours) were collected, manually assessed, and annotated using pre-defined ECG features. Using data from five out of seven centers (n = 634), two mortality prediction models were developed: (a) a logistic regression model using manually annotated ECG features, and (b) a pre-trained deep neural network (DNN) using the raw ECG waveforms. Data from two other centers (n = 248) were used for external validation. Performance of both prediction models was similar, with a mean area under the receiver operating curve of 0.69 [95%CI 0.55–0.82] for the logistic regression model and 0.71 [95%CI 0.59–0.81] for the DNN in the external validation cohort. After adjustment for age and sex, ventricular rate (OR 1.13 [95% CI 1.01–1.27] per 10 ms increase), right bundle branch block (3.26 [95% CI 1.15–9.50]), ST-depression (2.78 [95% CI 1.03–7.70]) and low QRS voltages (3.09 [95% CI 1.02-9.38]) remained as significant predictors for mortality.

Conclusion

This study shows that ECG-based prediction models at admission may be a valuable addition to the initial risk stratification in admitted COVID-19 patients. The DNN model showed similar performance to the logistic regression that needs time-consuming manual annotation. Several ECG features associated with mortality were identified.

Figure 1: 

Overview of methods, using and example case: (left) logistic regression and (right) deep learning. This specific case had a high probability of in-hospital mortality (above the threshold of 30%). Follow-up of this case showed that the patient had died during admission.

The incidence and clinical ramifications for leadless pacemaker fixation mechanism exposure on the epicardial surface

Funding Acknowledgements

Type of funding sources: Private company. Main funding source(s): Boston Scientific and Abbott

Background

Leadless pacemaker (LP) fixation mechanism exposure (FE) by penetration of the epicardial surface has been described. Previously reported animal model studies showed FE for 7/10 Micra LPs, versus 4/10 CapSureFix Novus RV pacing leads (both Medtronic). However, it is unknown whether FE causes pericardial effusion or pericarditis or does not have clinical significance.

Purpose

To determine the incidence of FE of a novel LP in a chronic animal model and its association with acute or chronic pericardial effusion.

Methods

 Canine subjects were implanted with novel LPs (Boston Scientific) in an ongoing study. Acute pericardial effusion was assessed by post-procedural transthoracic echocardiography (TTE). Chronic pericardial effusion was assessed by TTE 90 days after implantation and post-mortem assessed pericardial fluid colour (PFC) and volume (PFV). FE was assessed visually at necropsy. Mann-Whitney U tests and chi-squared tests were used to determine whether greater PFV, more haemorrhagic PFC or LP implantation location differed significantly between subjects with and without FE.

Results

 Results to date are reported. Canine subjects (n = 71) were chronically implanted with LPs. Due to 14 in-vivo retrievals, data is shown of 57 subjects with LPs in situ at necropsy. Pre-deployment radiocontrast injection confirmed LP position (RV apex n = 41; RV apicoseptal n = 16), and mechanical stability and electrical testing confirmed adequate talon fixation after deployment. Necropsy after median 94 days (IQR 91-540) demonstrated FE in 11 cases (19%) (figure). No acute nor chronic pericardial effusion was seen on TTE. Mean PFV for animals with and without FE was 1.8 and 1.6 cc, respectively. FE did not show an association with PFV or colour (p= 0.53 and p = 0.83, respectively). For two animals, PFV and PFC are not available; FE was not observed in either of these cases. LP implantation location was not associated with incidence of FE (p = 1.00).

Conclusion

 Fixation mechanism exposure by the talons of a novel leadless pacemaker was observed in 19% of animals implanted and was not associated with acute or chronic pericardial effusion. Abstract Figure 1

A comprehensive scoping review on transvenous temporary pacing therapy

Transvenous temporary cardiac pacing therapy (TV-TP) is widely used to treat life-threatening arrhythmias. Yet aggregated evidence on TV-TP is limited. We conducted a systematic scoping review to evaluate indications, access routes and complications of TV-TP, as well as permanent pacemaker therapy (PPM) following TV-TP. Clinical studies concerning TV-TP were identified in Ovid MEDLINE. Case studies and studies lacking complication rates were excluded. To assess complication incidence over time, differences in mean complication rates between 10-year intervals since the introduction of TV-TP were evaluated. We identified 1398 studies, of which 32 were included, effectively including 4546 patients. Indications varied considerably; however TV-TP was most commonly performed in atrioventricular block (62.7%). The preferred site of access was the femoral vein (47.2%). The mean complication rate was 36.7%, of which 10.2% were considered serious. The incidence of complications decreased significantly between 10-year interval groups, but remained high in the most recent time period (22.9%) (analysis of variance; p < 0.001). PPM was required in 64.2% of cases following TV-TP. Atrioventricular block was the primary indication for TV-TP; however indications varied widely. The femoral vein was the most frequent approach. Complications are common in patients undergoing TV-TP. Although a decrease has been observed since its introduction, the clinical burden remains significant. The majority of patients who underwent TV-TP required PPM therapy.

The modular cardiac rhythm management system: the EMPOWER leadless pacemaker and the EMBLEM subcutaneous ICD

Cardiac implantable electronic devices have been successfully treating patients with brady- and tachyarrhythmias for decades. However, there are still significant complications related to this therapy modality, many related to the transvenous lead. Paradigm-shifting technologies, such as the subcutaneous implantable cardioverter-defibrillator (S-ICD) and leadless cardiac pacemakers (LCP), have emerged to address these complications. The novel modular cardiac rhythm management (mCRM) system, consisting of a communicating antitachycardia pacing-enabled LCP and S‑ICD, is the first system to integrate wireless intrabody communication between devices to allow for coordination of leadless pacing and defibrillator therapy delivery. In this review, the design and concept of the mCRM system are presented and available evidence is summarized.